Reading mode: Light Dark

More Moire

Go to last post
djbrom said:
Seeing moire wouldnt really make me think that it had the capability of being very sharp.
Click to expand...
Well, my own experience is that when I've seen moire quite often with a certain camera, it turns out that it is indeed capable of very sharp results most of the time. But with certain other cameras that never quite give the same level of sharpness (that mostly I'd like), then they rarely show moire. So in some ways, it can be a fairly reliable way of determining the sharpness.

However its not always quite that simple. For example, I'd say the D300 and D3 seem to have a pretty similar level of per pixel sharpness, yet I see moire now and again on the D300, yet I don't with the D3 shooting the same things at the same time in nearly identical conditions. I don't have enough information on the AA filters to be able to work out why this is the case, but observation over many hundreds of shoots and hundreds of thousands of shots tells me its the case. I find it fascinating and its one of the reasons I like to discuss these things.
 
I would suggest that it is possible the some people perceived 'oly colour' is because Olympus was trying to create the best image possible by getting the best light possible to arrive at the sensor.

Fixing artifacts creates subtle colour shifts. The DWIA site notes that the colour of the E5 has fallen off compared to the E3. Now, there could be many reasons for this, but it is possible that by allowing more optical artifacts to be present and fixing them in software is causing colour shifts.

Did you see the purple casts in msusic's E5 crops that weren't present in his E3 crops? Even though it has been there since he first posted them, no one noticed. To me, that is evidence of artifact tolerance.

What do you think?
 
On my E-5 I saw Moire like this, maybe worse from the tile roof of a house. It only showed up in the Lightroom thumbnail image. After opening the image, it was gone. My guess is that LR used some processing to remove it. Since my images were for a client, I was very worried for a second but after opening my heart went back to normal.
--
One great 12mp image is worth millions of mediocre 24mp ones...
 
rovingtim said:
I would suggest that it is possible the some people perceived 'oly colour' is because Olympus was trying to create the best image possible by getting the best light possible to arrive at the sensor.

Fixing artifacts creates subtle colour shifts. The DWIA site notes that the colour of the E5 has fallen off compared to the E3. Now, there could be many reasons for this, but it is possible that by allowing more optical artifacts to be present and fixing them in software is causing colour shifts.

Did you see the purple casts in msusic's E5 crops that weren't present in his E3 crops? Even though it has been there since he first posted them, no one noticed. To me, that is evidence of artifact tolerance.

What do you think?
Click to expand...
Might it simply be a matter of the software not being "properly configured" to the hardware? As boggis' demonstration with four different RAW conversions demonstrated:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38561969

"inadequacies" in the photo may be more a matter of software than hardware.
 
Great Bustard said:
Might it simply be a matter of the software not being "properly configured" to the hardware? As boggis' demonstration with four different RAW conversions demonstrated:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38561969

"inadequacies" in the photo may be more a matter of software than hardware.
Click to expand...
Sure, different software is more effective than others. How not?

I would suggest that software correction is like AF: some are better than others but none are perfect.

I don't believe any software can be totally accurate in determining what is moire and artifact and what is detail. If this is true, wherever it gets it wrong it will either kill detail or colour or create colour or false detail. How not?

That is why Olympus believed optical correction was the way to go. It didn't require a computer to determine what was image and what was artifact and had the potential to create cleaner files.

However, all this is determinant on people's tolerance to artifacts. If no one is noticing the artifacts and everyone is noticing the sharpness, then the way is clear for the manufacturers.
 
rovingtim said:
Great Bustard said:
Might it simply be a matter of the software not being "properly configured" to the hardware? As boggis' demonstration with four different RAW conversions demonstrated:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38561969

"inadequacies" in the photo may be more a matter of software than hardware.
Click to expand...
Sure, different software is more effective than others. How not?

I would suggest that software correction is like AF: some are better than others but none are perfect.

I don't believe any software can be totally accurate in determining what is moire and artifact and what is detail. If this is true, wherever it gets it wrong it will either kill detail or colour or create colour or false detail. How not?
Click to expand...
The Bayer CFA itself is subject to "getting it wrong" on occasion.
rovingtim said:
That is why Olympus believed optical correction was the way to go. It didn't require a computer to determine what was image and what was artifact and had the potential to create cleaner files.
Click to expand...
I wouldn't call a stronger AA filter "optical correction" -- I would say "optical design choice ". That choice, of course, is the balance between demosaicing errors and captured detail.

That is, a strength of the AA filter simple trades one error for another. Where the best balance lies depends, obviously, not only on the QT (quality threshold) of the viewer, but the types of pics they take.

The manufacturer, of course, has to come up with a "one size fits all" solution, and, as we all know, one size never fits all.

Interestingly, there is another angle. By using the same mechanism for sensor IS, a camera could be designed so that the shake of the sensor during the exposure acts as the AA filter, thus making a firmware option for "AA filter strength" user controllable. Another plus for sensor IS, eh?
rovingtim said:
However, all this is determinant on people's tolerance to artifacts. If no one is noticing the artifacts and everyone is noticing the sharpness, then the way is clear for the manufacturers.
Click to expand...
Here's another intermediary solution that many might like. Let's say you love the 14-35 / 2 on the E3 due to its sharpness, but moire is sometimes an issue. Well, if you use a less sharp lens on the E5, say a 14-54 / 2.8, the blur from the lens will act as an AA filter, and you'd get 14-35 / 2 sharpness on an E3 with a 14-54 / 2.8 on an E5, and the same level of artifacts.

Thus, the E5 + 14-54 / 2.8 solution delivers the same results as the E3 + 14-35 / 2 combo with less size, weight, and cost!

It's not unlike the Equivalence argument of less expensive and using slower lenses on a larger sensor vs more expensive and faster lenses on a smaller sensor.

Something to consider (and test).
 
Great Bustard said:
The Bayer CFA itself is subject to "getting it wrong" on occasion.
Click to expand...
Indeed, but unless I'm misunderstanding something, that is something a well-tuned AA can mitigate.
Great Bustard said:
I wouldn't call a stronger AA filter "optical correction" -- I would say "optical design choice ". That choice, of course, is the balance between demosaicing errors and captured detail.
Click to expand...
When I said optically correct light, while I was including the AA, I meant the entire optical line.
Great Bustard said:
That is, a strength of the AA filter simple trades one error for another.
Click to expand...
I think perhaps we could both use a little more information here. My understanding is that AA isn't simply weaker or strong, but is a very complex optical element. That is why it costs so much.
Great Bustard said:
The manufacturer, of course, has to come up with a "one size fits all" solution, and, as we all know, one size never fits all.
Click to expand...
Indeed.
Great Bustard said:
Here's another intermediary solution that many might like. Let's say you love the 14-35 / 2 on the E3 due to its sharpness, but moire is sometimes an issue. Well, if you use a less sharp lens on the E5, say a 14-54 / 2.8, the blur from the lens will act as an AA filter, and you'd get 14-35 / 2 sharpness on an E3 with a 14-54 / 2.8 on an E5, and the same level of artifacts.

Thus, the E5 + 14-54 / 2.8 solution delivers the same results as the E3 + 14-35 / 2 combo with less size, weight, and cost!
Click to expand...
An interesting argument, but then you have to tune your lenses to a particular sensor. I think it might be more flexible to a lens system to put in a well-tuned AA in from of their sensors and then put really good lenses in front.

Unless I misunderstand lens design, contrast, bokeh and sharpness are lens design trade-offs. It seems to me that most are now measuring lenses solely on sharpness which is only part of IQ. That is why I think sharpness has become the new holy grail over all other things.

Going to bed now.
 
Great Bustard said:
rovingtim said:
I would suggest that it is possible the some people perceived 'oly colour' is because Olympus was trying to create the best image possible by getting the best light possible to arrive at the sensor.

Fixing artifacts creates subtle colour shifts. The DWIA site notes that the colour of the E5 has fallen off compared to the E3. Now, there could be many reasons for this, but it is possible that by allowing more optical artifacts to be present and fixing them in software is causing colour shifts.

Did you see the purple casts in msusic's E5 crops that weren't present in his E3 crops? Even though it has been there since he first posted them, no one noticed. To me, that is evidence of artifact tolerance.

What do you think?
Click to expand...
Might it simply be a matter of the software not being "properly configured" to the hardware? As boggis' demonstration with four different RAW conversions demonstrated:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38561969

"inadequacies" in the photo may be more a matter of software than hardware.
Click to expand...
The principle problem is that you can't assess the raw capture -- in order to get an image the raw file must be processed.

From my testing it appears that Olympus Viewer is preventing moire through a "software" AA process that also reduces detail. Most other raw developers yield significantly more detail, but you may get moire (including false colour in black text, in the case of ACR). Only the "amaze" demosaic algorithm, that RawTherapee can be set to use, seems to be able to get detail and avoid moire -- in fact, it yields the most detail.

The question is: does the moire exist in the raw file? If it does, then the E-5 is where the issue lies. If it doesn't, then it's the post-processing -- notably demosaicing. (It is possible that "amaze" has some clever algorithms for detecting moire and removes it, so it could actually be present.)

We don't really know how the E-5 system (light AA filter, "detail engine" to deal with moire) stacks up against the normal process. Most users report that moire isn't an issue, but this obviously depends on what you shoot.
 
rovingtim said:
Fixing artifacts creates subtle colour shifts. The DWIA site notes that the colour of the E5 has fallen off compared to the E3.
Click to expand...
What does "fallen off" mean? I have noticed that the E-5 has a more cyan tone in a blue sky to other Olympus cameras I have (especially the E-1), but that's a colour shift .
rovingtim said:
Now, there could be many reasons for this, but it is possible that by allowing more optical artifacts to be present and fixing them in software is causing colour shifts.
Click to expand...
I wouldn't think so. If you detect a false colour point then you would change it to an average of the surrounding colours. This shouldn't affect the colour balance of the photograph, even if you have a huge number of false-colour artefacts. A lighter AA filter should also introduce less possible colour interference, not more.
rovingtim said:
Did you see the purple casts in msusic's E5 crops that weren't present in his E3 crops? Even though it has been there since he first posted them, no one noticed. To me, that is evidence of artifact tolerance.
Click to expand...
Unless the shots were taken in the same light it is difficult to assess the colour shift. It could be that the E-3 colour was off. DPR seem to have stopped the colour accuracy testing using the GretagMacbeth ColorChecker chart, unfortunately.

Your default position seems to be that any difference between the E-3 and E-5 indicates a fault with the E-5. This isn't necessarily the case.
 
Fleabag said:
On my E-5 I saw Moire like this, maybe worse from the tile roof of a house. It only showed up in the Lightroom thumbnail image. After opening the image, it was gone.
Click to expand...
A thumbnail is usually a quick and dirty re-sampling, and will often produce artefacts.

Try opening the photo then scaling it down in small increments from 100%. You should see various artefacts occur at different scaling factors. This is to be expected.
 
boggis the cat said:
From my testing it appears that Olympus Viewer is preventing moire through a "software" AA process that also reduces detail. Most other raw developers yield significantly more detail, but you may get moire (including false colour in black text, in the case of ACR). Only the "amaze" demosaic algorithm, that RawTherapee can be set to use, seems to be able to get detail and avoid moire -- in fact, it yields the most detail.

The question is: does the moire exist in the raw file?
Click to expand...
you can open ORF in Irfan at 100% view and take a screen capture. This process apparently has no moire where it was found on the circular slide rule

--
Riley

any similarity to persons living or dead is coincidental and unintended
 
Doing a little research. About three quarters of the way down this page ...

http://www.luminous-landscape.com/reviews/cameras/14n-initial2.shtml

... there is an example of chromatic aberration from a Kodak (the windmill image) which the writer is concerned about and is later informed by his 'technically savvy' correspondents that this is colour aliasing as a result of a lack of anti-aliasing filter.

That is exactly the same thing I'm seeing in msusic's E5 crops vs the E3 crops.

Maybe instead of using the term moire, I should be using the term colour aliasing.
 
Rriley said:
boggis the cat said:
From my testing it appears that Olympus Viewer is preventing moire through a "software" AA process that also reduces detail. Most other raw developers yield significantly more detail, but you may get moire (including false colour in black text, in the case of ACR). Only the "amaze" demosaic algorithm, that RawTherapee can be set to use, seems to be able to get detail and avoid moire -- in fact, it yields the most detail.

The question is: does the moire exist in the raw file?
Click to expand...
you can open ORF in Irfan at 100% view and take a screen capture. This process apparently has no moire where it was found on the circular slide rule
Click to expand...
This is also the case for FastStone Image Viewer -- I use both IrfanView and FastStone -- because what they open is the embedded JPEG. The embedded JPEG is, of course, processed by Olympus' routines and so does not exhibit moire.
 
rovingtim said:
This was taken from the DIWA website

http://www.diwa-awards.net/wip4/test_result_overview.epl?product=6414&id=181182

Even though the E5 colour sensitivity has gone up slightly relative to the E3, its colour fidelity has fallen.



Click to expand...
Yes, provided the uncertainty in those graphs is relatively small.

I compared the data for the E-5, E-3, E-30 and E-PL2. They are all roughly the same for colour sensitivity, while the E-30 stands out for inaccurate colour reproduction. This seems a bit odd, as I do not recall anyone mentioning this with the E-30.

Those differences may not be significant when the (unknown) uncertainty is considered, so it seems a stretch to claim a significant colour shift is caused by Olympus' moire reduction processing. I have noted that my E-5 seems to render blue skies more cyan than my other cameras -- the E-1 rendering being the most blue. (This may also be the case for blue water.)

Links to the Diwa data:

E-5: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=8228&id=561243

E-3: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=6414&id=181182

E-30: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=6613&id=383702

E-PL2: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=8303&id=572446
 
Just thought that you may find this interesting. Scroll down to the bottom of each of the E-3 and E-5 results and look at the "DxO Vignetting" charts.

Far be it from me to suggest that Diwa would make a gigantic error and fail to notice it, but... So what is the explanation for this?

Olympus' "detail engine" algorithms cause extreme vignetting -- fortunately this is, inexplicably, unnoticeable in use? ;)
 
Gidday Cat
boggis the cat said:
rovingtim said:
This was taken from the DIWA website

http://www.diwa-awards.net/wip4/test_result_overview.epl?product=6414&id=181182

Even though the E5 colour sensitivity has gone up slightly relative to the E3, its colour fidelity has fallen.

dpr : galleries/356229571/photos/1174134
Click to expand...
Yes, provided the uncertainty in those graphs is relatively small.

I compared the data for the E-5, E-3, E-30 and E-PL2. They are all roughly the same for colour sensitivity, while the E-30 stands out for inaccurate colour reproduction. This seems a bit odd, as I do not recall anyone mentioning this with the E-30.
Click to expand...
Mate, there have been lots of comments about the colour shift and fixes for the E-30. I have adjusted the colour by setting the PICTURE MODE to NATURAL, and SATURATION to -2. Others have changed the WB settings for magenta and green axes (IIRC ... ).
boggis the cat said:
Those differences may not be significant when the (unknown) uncertainty is considered, so it seems a stretch to claim a significant colour shift is caused by Olympus' moire reduction processing. I have noted that my E-5 seems to render blue skies more cyan than my other cameras -- the E-1 rendering being the most blue. (This may also be the case for blue water.)

Links to the Diwa data:

E-5: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=8228&id=561243

E-3: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=6414&id=181182

E-30: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=6613&id=383702

E-PL2: http://www.diwa-labs.com/wip4/test_result_overview.epl?product=8303&id=572446
Click to expand...
Thanks for these links. I have added the E-30 to my favourites, and bookmarked your post for closer examination later. Interesting stuff.

--
Regards, john from Melbourne, Australia.
(see profile for current gear)
Please do not embed images from my web site without prior permission
I consider this to be a breach of my copyright.
-- -- --

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

Gallery: http://canopuscomputing.com.au/gallery2/main.php



Bird Control Officers on active service.

Member of UK (and abroad) Photo Safari Group
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top